It's been a solid two weeks since I posted for real on the blog. It
took a little restraint at first; every stray thought I had seemed
like the perfect topic for a blog post. Eventually I was able to free
my mind from the thrall of my cybernetic overlords. Also, I skied.
It was a nice time -- family, presents, fine wine, and the enduring joy
of repeatedly digging four-wheel-drive vehicles out of snowbanks. I
hope you also had nice holidays.
Now I'm back in the harness and will get back to you with some real
medical posts soon enough. For the moment, I will throw in my $0.02 on
a debate regarding the role of Mid-level providers (MLPs) in the ED,
specifically NP's and PA's. Scalpel has made his point in spades, , , , and , while Ten out of Ten has provided a counter-point , and Happy the Hospitalist chimed in .
Disclaimer -- I am not an expert in licensing requirements, liability,
or reimbursement other than in my particular state. Also, there is
variability in both the MD and MLP populations -- I've known PAs who
were exceptional clinicians, and MDs who were, we joked, "licensed to
kill." And vice versa. Your mileage may vary. I'm not interested in
a "PA's suck!" "No, doctors suck!" argument. Having said that, I will
dive in:
What is a PA or NP? First of all, it's important to understand that in most states, they are "Licensed Independent Providers,"
which means that they are qualified to examine and treat patients and
bill for their services without direct physician oversight. PAs
generally need a nominal supervising physician, and state laws vary as
to how close the supervision need be. In our state, the doc must be
physically on the premises except in critical access locations. NPs
have less restrictive requirements. Both NP and PA training programs
are highly competitive, and in most cases will accept only applicants
with significant healthcare experience (the consequence is that the MLP
ranks are full of former nurses, paramedics, and corpsmen). The
training can be as little as 2 to 4 years (not counting
prerequisites), and yields a masters-level degree. In my experience,
the intensity, depth and breadth of the training is substantially less
than that in medical school and residency, and the lack of standardized
post-graduate training for MLPs requires significant on-the-job
learning for new graduates.
The consequence of the more superficial education of MLPs is that they
are usually required to function within a narrowly defined scope of
practice. In an ER, that may be limited to minor traumatic injuries
and other simple complaints. I have known PAs who were highly
specialized as vascular or neurosurgery assistants; their understanding
of their field far exceeded my own, but they functioned as extenders of
their supervising docs and bore limited independent responsibility.
For an MLP, knowing your scope of practice and staying within it is
essential. (The same principle applies to physicians, I might add,
although our scope is comparatively expansive.)
How are MLPs utilized in ERs?
This is highly dependent on local and institutional issues and on the
experience and comfort level that a department may have with MLPs.
The most restrictive environments require the PA to present all cases
to a doc and require the doc to see the patient as well -- in essence,
this has the PA function like a resident physician. In other cases,
the PA just has to present the patient, with the doc electing to see or
not see them as they feel is indicated. Some EDs just have the docs
review and co-sign all the PA charts, and others have a QA process by
which a random sample of the PA charts are reviewed retrospectively.
The more autonomously the MLPs operate, the more efficient it is, but
that must be balanced by how well the scope of practice is adhered to
and how much risk there is that a MLP might get in over his or her head
with an unexpectedly complex patient. In a well- run ER, there is
ample opportunity (and no disincentive) for MLPs to consult with or
transfer care to a physician, as needed. My opinion is that with
experienced MLPs and a carefully selected patient population, it is
possible to safely run a fast track with completely independent MLPs.
Why are MLPs staffing ERs at all? The primary reason is economic, though ,
there is a shortage of qualified EM physicians which also is an
incentive for ERs to hire MLPs if their patient demographics make
sense. But the main reason is economic. Consider a PA working a site
where the volume is not terribly high -- 2.5 patients per hour. Fast
Track acuity typically translates to an average value of about 2.5 RVU
per patient. So the PA is bringing in 6.25 RVU/hour, which at a
conservative $40/RVU collection rate is $250/hour. Subtract $50 for
expenses and pay the PA $60/hour, and the remaining $140/hour is profit
for the employer. Incentivize your MLPs so they are a little faster,
and your profit margin only goes up. This is an effective subsidy to
the physician income base, one of the few refuges available in an era
of shrinking reimbursement.
Speaking of reimbursement, how does that work for MLPs?
That depends entirely on the internal policies of each individual
patient's payer. When a PA sees a patient, it is coded with the ICD-9
and CPT codes based on their documentation, just like when a doctor
sees a patient. However, some payers do and some do not issue
provider ID numbers to PAs. Medicare and most governmental payers do
issue provider IDs to PAs, and if the PA is the sole provider listed on
the billing form, will reduce the allowable fee by 15%. Most
commercial payers, in my experience, also credential PAs, and pay at
the same rate that they do for physician services, although some may
apply a random reduction in the allowable (read your contracts!)
between 5-20%. Medicare will pay an E/M code at the physician rate if
it is a shared service, meaning that there must be documentation that
the physician had (at a minimum) a face-to-face interaction with the
patient. However, procedure
codes, from lacerations up to and including Critical Care, may not be
shared services and will be paid at the rate of the provider who
actually performed the procedure. Thus if you think you can improve
reimbursement by documenting that you supervised a PA's laceration
repair, think again!
If the payer does not allow a charge to be billed in the PA's name,
then the charge will be issued in the name of the supervising physician
with the PA listed in the second position on the billing form (this
will be ignored by the payer, but is necessary for internal
record-keeping). Usually these get paid at 100% of the allowable.
Scalpel made a strange argument that MLP services should be at a steep
discount from physician rates. While this would be a great way to
eradicate MLPs from the health care landscape, I don't see much
validity to this. A service provided is a service provided, and the
worth of the service, performed competently, does not vary according to
the credentials of the individual who provides it. A laceration
repair is not worth more to the patient if the doc does it. A chest
tube pays the same whether the surgeon or I put it in. Scalpel and I
get paid the same though I am AEBM certified and he, apparently, is
not. As far as I can tell, the only rationale behind the 15% holdback
from CMS for MLP services is, "Because we can."
What about the vicarious liability implications of using MLPs in the ED?
Obviously, there is always liability, but it is generally low. MLPs
are under-represented in ED med-mal cases, and given the lower acuity
of the patients they see, that makes sense, as does the fact that in
most cases the dollar amount at stake is low. If you as the
supervising physician never saw the patient, then you can seek to have
your name dropped from the case due to absence of doctor-patient
relationship. You may be on the hook for negligent supervision, but
that is more commonly directed at your mutual employer. There are
occasional cases in which a doctor who never saw the patient is found
to have some responsibility, but that is more typically in cases where
something else happened (an unlicensed PA, or falsification of the
chart). In my experience (fortunately quite limited) it is fairly
uncommon for the doc to even be named, if they never laid eyes on the
patient.
What's the big picture?
Scalpel thinks we should "Just Say No to Fast Tracks," and he's partly
right. There is a significant added expense when minor ailments are
treated in the ER (which is to say that Fast Tracks are profit centers
for physicians and hospitals). Given that cognitive services are
undercompensated relative to
the real work and risk that go into them, it is essential for ER groups
to retain the simple cases to cross-subsidize the work on the complex,
sick patients. Also, given the ongoing collapse of primary care, it
is becoming progressively more difficult for patients to receive care
for acute illnesses and injuries at their doctors' offices, and the ER
is an attractive one-stop-shop for patients. You can be in and out in
a well-run Fast Track in 90 minutes. The macro-economic climate
ensures that the customer demand is there, and we do nobody any favors
by refusing to meet that demand.
In an ideal world, when primary care physicians are well-paid and
plentiful, perhaps that demand will cease to exist and the patients
will all go back to outpatient centers. Or, more likely, CMS will cut
reimbursement for Type B ED services (aka Fast Track) and hospitals
will no longer have an incentive to grow that service line. From a
queuing theory perspective, Fast Tracks (and the MLPs that run them)
are essential to clearing out the lower rungs of the acuity ladder.
It's short-sighted and vindictive to insist that less-urgent patients
must wait until the truly sick have been seen. Run a good fast track,
and everybody gets seen faster.
My personal opinion is that while I have nothing against PAs, in an
ideal world, I would not employ them. It makes the management of a
group more difficult to have two different classes of providers at
different wage scales, and there is an inevitable tension between the
two groups which I find is not conductive to good morale. The
economic argument, however, is very compelling, and the decrease in
income were we to change to an all-physician model would be painful
indeed. Besides, we are not building an organization from a clean
sheet of paper; we have had PAs for a couple of decades, they are good
friends and colleagues who deliver good care, and their place in our
organization is quite secure.